Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS5912824 A
Publication typeGrant
Application numberUS 08/932,642
Publication dateJun 15, 1999
Filing dateSep 17, 1997
Priority dateSep 18, 1996
Fee statusLapsed
Also published asDE69717382D1, DE69717382T2, EP0831407A2, EP0831407A3, EP0831407B1
Publication number08932642, 932642, US 5912824 A, US 5912824A, US-A-5912824, US5912824 A, US5912824A
InventorsKoichi Sawahata
Original AssigneeNec Corporation
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Ion implantation simulation method
US 5912824 A
Abstract
Disclosed is an ion implantation simulation method including calculating a particle scattering process of each of sample particles as a simulation target using a Monte Carlo method, determining that the particle has stopped when scattering calculation gives zero energy of the particle, continuously performing the scattering calculation when the energy is not 0. When the scattering calculation yields an energy of the particle that has decreased to α (0≦α≦1) times the energy value at the time of implantation, the particle Is divided into a predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division. The scattering calculation and the particle division process is repeated until particles having non-zero energy values are divided the predetermined number of times M (M is an integer) counting from the first division, and consequently, the weight of the particle becomes 1/NM that of the initially implanted particle, performing the scattering calculation for particles which have been divided the predetermined number of times M until the energy value becomes 0, and determining the positions where the energy values of all the sample particles become 0 as stop positions in a substrate, thereby obtaining an impurity distribution.
Images(4)
Previous page
Next page
Claims(2)
What I claim is:
1. An ion implantation simulation method comprising calculating a particle scattering process of each of sample particles as a simulation target using a Monte Carlo method, determining that the particle has stopped when the scattering calculation yields zero energy of the particle, continuously performing the scattering calculation when the energy is not 0, when the scattering calculation yields an energy of the particle that has decreased to α (0≦α≦1) times the energy value at the time of implantation, dividing the particle divided into a predetermined number N (N is an integer) such that a weight of the particle after division becomes 1/N that before division, further calculating the scattering process for the particle after division, determining that the particle has stopped when the energy becomes 0. when the energy is not 0 and decreases to α (0≦α≦1) times the energy value in the previous division, the particle is further divided into a predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division, further calculating the scattering process for the particle after division, repeatedly the scattering calculation and the particle division process until particles having non-zero energy values are divided the predetermined number of times M (M is an integer) counting from the first division, and consequently, the weight of the particle becomes 1/NM that of the initially implanted particle, performing the scattering calculation for particles which have been divided the predetermined number of times M until the energy value becomes 0, and determining positions where the energy values of all the sample particles become 0 as stop positions in a substrate, thereby obtaining an impurity distribution.
2. A method according to claim 1, wherein, when an energy loss due to an electronic or nuclear stopping power is larger than a binding energy of substrate atoms at a position where the particle collides against the atom, the atom forced out upon collision is also included in the particle and scattering calculation on the basis of monte Carlo Method is performed.
Description
BACKGROUND OF THE INVENTION

1. FIELD OF THE INVENTION

The present invention relates to ion implantation for impurity doping in a semiconductor manufacturing process and, more particularly, to an ion implantation simulation method Of calculating a distribution of impurity particles which are ion-implanted into a substrate and stop at certain positions because of the energy loss after repetitive collision against the substrate.

2. DESCRIPTION OF THE PRIOR ART

Ion implantation simulation using a Monte Carlo method (to be referred to as "Monte Carlo ion implantation simulation" hereinafter) is described in Ryo Dan, "Process Device Simulation Technology", p. 60. In this ion implantation simulation, a process that implanted ions scatter and lose energy while colliding against atomic nuclei and electrons in the substrate is calculated using a probabilistic technique. More specifically, a random number is generated for every collision process to determine the relative position with respect to a target atom, i.e., a collision parameter. Scattering (energy transition and direction) of the implanted ions is calculated on the basis of the collision parameter, thereby obtaining the distribution of impurity particles which have finally stopped in the substrate. In this specification, calculation of the scattering process of implanted ions using the above-described Monte Carlo method will be referred to as "scattering calculation" hereinafter.

To accurately calculate the distribution using this simulation method, a lot of trajectories of implanted ions must be calculated, resulting in a very long calculation time. A solution to this problem is described in S. H. Yang et al., "A More Efficient Approach for Monte Carlo Simulation of Deeply-Channeled Implanted Profiles in Single-Crystal Silicon", NUPAD V, pp. 97-100, (1994).

According to this technique, Monte Carlo ion implantation simulation is performed first using a certain number of sample particles to obtain an impurity profile as shown in FIG. 1A. Next, with reference to the resultant profile, positions in the depth direction where the sample particles are divided are determined. These positions are represented by d1, d2, and d3 in FIG. 1B. Sample particles which have reached the depths d1, d2, and d3 are divided, and the Monte Carlo ion implantation simulation is performed again. With this process, a profile whose tail portion has minimum noise can be obtained, as shown in FIG. 1B.

In the conventional simulation method, however, simulation can hardly be extended to two or three dimensions. The reason for this is as follows. To extend simulation to two or three dimensions, a function of determining a two- or three-dimensional line segment sequence or plane is required to divide sample particles.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a Monte Carlo ion implantation simulation method which allows more accurate ion implantation simulation not only for a one-dimensional structure but also for a two- or three-dimensional structure without prolonging the calculation time.

In order to achieve the above object, according to the present invention, there is provided an ion implantation simulation method comprising calculating a particle scattering process of each of sample particles as a simulation target using a Monte Carlo method, determining that the particle has stopped when scattering calculation yields zero energy of the particle, continuously performing the scattering calculation when the energy is not 0, when the scattering calculation yields an energy of the particle that has decreased to α (0≦α≦1) times the energy value at the time of implantation, dividing the particle divided into a predetermined number N (N is an integer) such that a weight of the particle after division becomes 1/N that before division, further calculating the scattering process for the particle after division, determining that the particle has stopped when the energy becomes 0, when the energy is not 0 and decreases to a (0≦α≦1) times the energy value in the previous division, the particle is further divided into a predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division, further calculating the scattering process for the particle after division, repeatedly the scattering calculation and the particle division process until particles having non-zero energy values are divided the predetermined number of times M (M is an integer) counting from the first division, and consequently, the weight of the particle becomes 1/NM that of the initially implanted particle, performing the scattering calculation for particles which have been divided the predetermined number of times M until the energy value becomes 0, and determining positions where the energy values of all the sample particles become 0 as stop positions in a substrate, thereby obtaining an impurity distribution.

According to this method, when the Monte Carlo ion implantation simulation is to be performed, particles are divided with reference to the degree of drop in particle energy, so positions where the particles are divided need not be determined in advance. Therefore, the calculation accuracy can be increased without increasing the calculation time. In addition, the simulation method of the present invention can be easily applied to simulation for a two- or three-dimensional structure.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a graph showing a profile obtained by performing conventional Monte Carlo ion implantation simulation without division for a predetermined number of particles;

FIG. 1B is a graph showing a profile obtained by performing simulation while dividing the particles at predetermined depths on the basis of the profile obtained in FIG. 1A;

FIG. 2 is a flow chart for explaining a Monte Carlo ion implantation simulation method according to the first embodiment of the present invention;

FIG. 3A is a graph showing a profile obtained by performing the Monte Carlo ion implantation simulation of the present invention for 1,000 implanted sample particles;

FIG. 3B is a graph showing a profile obtained by performing the conventional Monte Carlo ion implantation simulation without division for 2,500 implanted sample particles; and

FIG. 4 is a flow chart for explaining a Monte Carlo ion implantation simulation method according to the second embodiment of the present invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

The present invention will be described below with reference to the accompanying drawings.

In the Monte Carlo ion implantation simulation of the present invention, the scattering process of each sample particle as a simulation target is calculated using a Monte Carlo method. When the scattering calculation gives zero energy of the particle, it is determined that the particle has stopped. Otherwise, the scattering calculation is continued. When the scattering calculation yields an energy of the particle that has decreased to α (0≦α≦1) times the energy value at the time of implantation, the particle is divided into a predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division. The scattering process of the particle after division is calculated. When the energy becomes 0, it is determined that the particle has stopped. Otherwise, when the energy decreases to α (0≦α≦1) times the energy value in the previous division, the particle is further divided into a predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division. The scattering process is further calculated for the particle after division. In this way, the scattering calculation and the particle division process are repeated until particles having non-zero energy values are divided a predetermined number of times M (M is an integer) counting from the first division operation, and consequently, the weight of the particle becomes 1/NM that of an initially implanted particle. For particles which have been divided the predetermined number of times M, the scattering calculation is performed until the energy value becomes 0. The positions where the energy values of all implanted particles become 0 are regarded as stop positions in the substrate, thereby calculating the impurity distribution.

An ion implantation simulation method according to the first embodiment of the present invention will be described next in detail with reference to the flow chart shown in FIG. 2. In this embodiment, assume that the simulation is done using a computer.

At the start of simulation, simulation setting conditions are input. More specifically, the ion implantation energy, the dose, and the like are input as ion implantation conditions. The number of sample particles as a simulation target, the division frequency M (M is an integer) defining the number of times of division for one particle, the division number N (N is an integer) defining the number of particles obtained by dividing one particle, the division reference multiplier α which is set to execute particle division with reference to a decrease in particle energy value, i.e., to execute particle division when the particle energy value becomes α (0≦α≦1) times that in the previous division, and the like are input as analysis conditions (step P1).

Assuming one sample particle is implanted, scattering calculation is started (step P2). The number of times of division execution is represented by m (0≦m≦M). The energy of a particle whose division frequency is m is represented by Em, and an initial energy E0 of a particle whose division frequency is 0 is set as an implantation energy. The energy of a particle at an arbitrary position is represented by ε.

A random number is generated, and the position of an atom where the implanted particle will collide next is calculated using the Monte Carlo method (step P3). The energy lose due to the electronic or nuclear stopping power when the implanted sample particle moves to the position where it collides against the atom is calculated on the basis of the calculation in step P3, and the energy ε of the particle is updated on the basis of the calculation (step P4).

Next, it is determined whether the energy ε of the particle, which is updated in step P4, has become 0 (step P5). If YES in step P5, it is determined that the particle has stopped, and the flow advances to step P6. If NO in step P5, the flow advances to step P8 to continuously calculate the trajectory of the particle.

In step P8, it is determined whether the number of times of executed division has reached the predetermined frequency M and whether the energy ε of the particle is α times the energy Em (or the initial energy E0) in the previous division. More specifically, if inequalities (1) below do not hold, the flow returns to step P3 to continue particle scattering calculation on the basis of the Monte Carlo method:

m<M, and ε<αEm                          (1)

This routine is repeated until the energy ε of the particle becomes 0, or inequalities (1) above are satisfied.

If it is determined in step P8 that inequalities (1) hold, m=m+1. The particle is divided into the predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division. The flow returns to step P3 to continue the scattering calculation for one of the divided particles. At this time, data of the position, the division frequency M, the energy value Em, the direction, and the weight of each of the N divided particles, except the particle for which calculation is continued, are stored in a predetermined stack area (step P10)

A case wherein it is determined in step P5 that the energy ε of the particle is 0 will be described. When the scattering calculation (or the scattering calculation and particle division process) yields zero energy ε of the sample particle as the calculation target, the stack area where data of the divided particles are stored in step P10 is accessed to determine whether scattering calculation for all particles divided in step P10 is ended (step P6). If NO in step P6, one of the particles stored in the stack area, for which calculation is not ended, is extracted. The flow returns to step P3 to start the scattering calculation on the basis of stored data.

If YES in step P6, it is determined whether calculation for all sample particles is ended (step P7). If NO in step P7, the flow returns to step P2 to start calculation for a particle extracted from the particles for which calculation is not ended.

In this manner, the scattering calculation is performed for all the divided particles. The scattering calculation and particle division process are repeated until particles having non-zero energy values are divided the predetermined number of times M (M is an integer) counting from the first division, and consequently, the weight of the particle becomes 1/NM that of an initially implanted particle. For particles which have been divided the predetermined number of times M, scattering calculation is performed until the energy value becomes 0. The positions where the energy values of all sample particles become 0 are regarded as stop positions in the substrate, thereby obtaining an impurity distribution.

FIGS. 3A and 3B show actual calculation examples. In FIG. 3A, the dotted line indicates a result obtained by performing the calculation according to the method of the present invention for 1,000 implanted sample particles. As the conditions for calculation, the division frequency M is 4, the number N of particles obtained by one division is 2, and the division reference multiplier α is 0.5. The number of particles finally becomes 24 =16 times, i.e., 16,000. Therefore, the calculation takes a longer time than that for calculation without division for 1,000 implanted particles. The solid line in FIG. 3A indicates a calculation result obtained by performing the conventional Monte Carlo ion implantation simulation without division for 100,000 implanted particles.

In FIG. 3B, the dotted line indicates a calculation result obtained by performing the conventional Monte Carlo ion implantation simulation without division for 2,500 implanted sample particles. The solid line in FIG. 3B indicates a calculation result obtained by performing the conventional Monte Carlo ion implantation simulation without division for 100,000 implanted sample particles. The time required to perform calculation for 1,000 particles using of the method of the present invention is almost the same as that required to perform calculation for 2,500 sample particles using the conventional method without division, so the two results are compared with the result of calculation for 100,000 particles.

When the dotted line in FIG. 3A is compared with that in FIG. 3B, calculation according to the present invention yields a more accurate result than that obtained by performing calculation without division for 2,500 sample particles.

The second embodiment of the present invention will be described with reference to the flow chart shown in FIG. 4. In this embodiment, ions are implanted into a silicon substrate through, e.g., an oxide film, and when the distribution of oxygen atoms which are forced from the oxide film into the silicon substrate by the implanted particles is to be calculated, the calculation accuracy is increased using the method of the present invention.

At the start of simulation, simulation setting conditions are input. More specifically, the ion implantation energy, the dose, and the like are input as ion implantation conditions. The number of sample particles as a simulation target, a division frequency M (M is an integer) defining the number of times of division for one particle, a division number N (N is an integer) defining the number of particles obtained by dividing one particle, a division reference magnification α which is set to execute particle division with reference to a decrease in particle energy value, i.e., to execute particle division when the particle energy value becomes α (0≦α≦1) times that in the previous division, and the like are input as analysis conditions (step Q1).

Assuming one sample particle is implanted, scattering calculation is started (step Q2). The number of times of division execution is represented by m (0≦m≦M). The energy of a particle whose division frequency is m is represented by Em, and an initial energy E0 of a particle whose division frequency is 0 is set as an implantation energy. The energy of a particle at an arbitrary position is represented by ε.

A random number is generated, and the position of an atom where the implanted particle will collide next is calculated using the Monte Carlo method (step Q3). The energy loss due to the electronic or nuclear stopping power when the implanted sample particle moves to the position where it collides against the atom is calculated on the basis of the calculation in step Q3, and the energy ε of the particle is updated on the basis of the calculation (step Q4)

In step Q5, it is determined whether the energy loss due to collision is larger than the binding energy of substrate atoms. More specifically, it is determined whether substrate atoms are forced out upon collision. If YES in step Q5, data of the type, energy, and moving direction of atoms forced out upon collision are stored in the stack area in step Q9. The energy of the atom at the division frequency m is represented by em, and an energy e0 that the atom forced out by the particle whose division frequency is 0 has acquired in step Q2 is represented by E0. These data are also stored in the stack area, and the flow advances to step Q6. In step Q6 and following steps, the atom forced out upon collision is also included in the particle referred in step Q6 and scattering calculation is performed for the particle included the atom.

If NO in step Q5, the flow directly advances to stop Q6.

Next, it is determined whether the energy ε of the particle, which is updated in step Q4, has become 0 (step Q6). If YES in step Q6, it is determined that the particle has stopped, and the flow advances to step Q7. If NO in step Q6, the flow advances to step Q10 to continuously calculate the trajectory of the particle.

In step Q10, it is determined whether the number of times of executed division has reached the predetermined frequency M and whether the energy ε of the particle is a times the energy Em (or the initial energy E0) in the previous division (step Q10) More specifically, if inequalities (2) below do not hold, the flow returns to step Q3 to continue particle scattering calculation on the basis of the Monte Carlo method:

m<M, and ε<αEm                          (2)

This routine is repeated until the energy ε of the particle becomes 0, or inequalities (2) above are satisfied.

If it is determined in step Q10 that inequalities (2) hold, m=m+1. The particle is divided into the predetermined number N (N is an integer) such that the weight of the particle after division becomes 1/N that before division. The flow returns to step Q3 to continue the scattering calculation for one of the divided particles. At this time, data of the position, the division frequency M, the energy value Em, the direction, and the weight of each of the N divided particles, except the particle for which calculation is continued, are stored in a predetermined stack area (step Q12).

A case wherein it is determined in step Q6 that the energy ε of the particle is 0 will be described. When the scattering calculation (or the scattering calculation and particle division process) yields zero energy ε of the sample particle as the calculation target, the stack area where data of the divided particles are stored in step Q12 is accessed to determine whether scattering calculation for all particles divided in step Q12 is ended (step Q7). If NO in step Q7, one of the particles stored in the stack area, for which calculation is not ended, is extracted. The flow returns to step Q3 to start the scattering calculation on the basis of stored data.

If YES in step Q7, it is determined whether calculation for all sample particles is ended (step Q8). If NO in step Q8, the flow returns to step Q2 to start calculation for a particle extracted from the particles for which calculation is not ended.

In this manner, the scattering calculation is performed for all the divided particles. The scattering calculation and particle division process are repeated until particles having non-zero energy values are divided the predetermined number of times M (M is an integer) counting from the first division, and consequently, the weight of the particle becomes 1/NM that of an initially implanted particle. For particles which have been divided the predetermined number of times M, scattering calculation is performed until the energy value becomes 0. The positions where the energy values of all sample particles become 0 are regarded as stop positions in the substrate, thereby obtaining an impurity distribution.

With the above method, not only implanted particles but also atoms forced out from the substrate can also be divided to increase the calculation accuracy.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US3765956 *Oct 19, 1971Oct 16, 1973C LiSolid-state device
US4136435 *Jan 31, 1977Jan 30, 1979Li Chou HMethod for making solid-state device
US4371406 *Jan 29, 1979Feb 1, 1983Li Chou HSolid-state device
US5420049 *Sep 9, 1993May 30, 1995The United States Of America As Represented By The Secretary Of The NavyMethod of controlling photoemission from porous silicon using ion implantation
US5737250 *Oct 12, 1995Apr 7, 1998Nec CorporationMethod and system for simulating ion implantation for performing simulation with avoiding overflow by adjusting memory consuming amount
US5787269 *Sep 19, 1995Jul 28, 1998Ricoh Company, Ltd.Process simulation apparatus and method for selecting an optimum simulation model for a semiconductor manufacturing process
Non-Patent Citations
Reference
1Dan; "Process Device Simulation Technology"; 1988; pp. 60-62; Sangyo Tosho Kabushiki Gaisha.
2 *Dan; Process Device Simulation Technology ; 1988; pp. 60 62; Sangyo Tosho Kabushiki Gaisha.
3Yang et al; "A More Efficient Approach for Monte Carlo Simulation of Deeply-Channeled Implanted Profiles in SIngle-Crystal Silicon"; 1994; pp. 97-100; IEEE.
4 *Yang et al; A More Efficient Approach for Monte Carlo Simulation of Deeply Channeled Implanted Profiles in SIngle Crystal Silicon ; 1994; pp. 97 100; IEEE.
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US6681201 *May 22, 2000Jan 20, 2004Nec Electronics CorporationSimulation method of breakdown
US6735556Jun 15, 2001May 11, 2004International Business Machines CorporationReal-time model evaluation
US7197437 *Sep 12, 2001Mar 27, 2007Kabushiki Kaisha ToshibaMethod, apparatus, and computer program for the Monte Carlo ion implantation simulation, and semiconductor device manufacturing method based on the simulation
US20120046924 *Jul 26, 2011Feb 23, 2012Kabushiki Kaisha ToshibaIon implanting simulating method and a computer-readable medium
Classifications
U.S. Classification703/6, 703/5
International ClassificationG06F19/00, G06Q50/00, G06Q50/04, H01L21/00, G06F17/50, H01L21/265
Cooperative ClassificationG06F2217/10, G06F17/5018
European ClassificationG06F17/50C2
Legal Events
DateCodeEventDescription
Aug 2, 2011FPExpired due to failure to pay maintenance fee
Effective date: 20110615
Jun 15, 2011LAPSLapse for failure to pay maintenance fees
Jan 17, 2011REMIMaintenance fee reminder mailed
Nov 17, 2006FPAYFee payment
Year of fee payment: 8
Feb 25, 2003ASAssignment
Owner name: NEC ELECTRONICS CORPORATION, JAPAN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:NEC CORPORATION;REEL/FRAME:013751/0721
Effective date: 20021101
Nov 22, 2002FPAYFee payment
Year of fee payment: 4
Sep 17, 1997ASAssignment
Owner name: NEC CORPORATION, JAPAN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SAWAHATA, KOICHI;REEL/FRAME:008724/0570
Effective date: 19970912